I am building a simulation model of human arm for analyzing its motion using the data obtained from sensors
I have a couple of questions with regard to the simulation model.
My sensor outputs a total of 10,000 data points in total and I would have to run all those data points through the simulation model to see if the results match.
But, It takes about 20 minutes to run through just 100 data points.
Is there a way to run through all the 10000 data points at a faster pace? Right now, it takes 3 hours to run 2000 data points.
By reducing the nstep, would the simulation model still go through all the motion data or is there any other way?
My second question is that while simulating the data points, sometimes the model gives the following error:
Unexpected exception in the library OOSol :
I understand that we correct this by increasing the string mesh. But, I tried removing the data points at which this error occurs. The model was able to run without any error after removing that specific data point which is causing the above-mentioned error.
But, I will have to remove a data point every time the error occurs and start the simulation from scratch. Like mentioned before, the simulation already takes 3 hours to run through once. Is there a way to tell the simulation model to automatically skip the data points causing this error and move forward with the next data point or is there a way to figure out which data points will cause this error so that we can remove all those data points even before we run through the simulation?
Please let me know. This would save a lot of time for me to get the output.
I posted a query in regards to managing the higher number data points and figuring out OOsol exception error before running it in the Anybody forum. Please refer to full post above for more information.
Would someone be able to help me with this problem?
Any suggestions would be highly appreciated.
Thanks in advance.
I don't know what you are analysing so it is hard to give you concrete advice. I don't think there are many ways to make it run faster directly inside AnyBody's GUI application.
Maybe you could split the analysis into multiple trials that each runs part of the simulation. That way you could parallelize the computation and then stitch the results back together in a post-processing step. That, of course, means that your would have to automate the simulations. The AnyPyTools that lets you do something like this from Python, but you could also automate AnyBody from Matlab if you have the more Matlab skills.
The OOSol error comes solver used for muscle wrapping. (I think). Those kind of errors often occur in extreme postures for which the model wasn't originally designed. Or it occurs if you reduce the resolution of your simulation (i.e. lowers nStep).
Is there a way to tell the simulation model to automatically skip the data points causing this error and move forward with the next data point or is there a way to figure out which data points will cause this error so that we can remove all those data points even before we run through the simulation?
I don't think there is a way to do this. Maybe you can run the simulation without muscles first and figure out which part of the simulation you are interested in. But I don't know your research question so maybe that is not an option.